From 7bc33979432abc3c142b3d803837fab0deab50ad Mon Sep 17 00:00:00 2001
From: Vojtech Forejt <forejtv@gmail.com>
Date: Sun, 4 Mar 2012 19:13:31 +0000
Subject: [PATCH] text file filler, first version, untested

git-svn-id: https://www.prismmodelchecker.org/svn/prism/prism/trunk@4776 bbc10eb1-c90d-0410-af57-cb519fbb1720
---
 prism/etc/README.txt              |   5 +
 prism/etc/scripts/prism-filler.py | 187 ++++++++++++++++++++++++++++++
 2 files changed, 192 insertions(+)
 create mode 100755 prism/etc/scripts/prism-filler.py

diff --git a/prism/etc/README.txt b/prism/etc/README.txt
index fe6df656..15476b7a 100644
--- a/prism/etc/README.txt
+++ b/prism/etc/README.txt
@@ -11,6 +11,11 @@ Where appropriate, comments/instructions are included in the files themselves.
    - prism3to4 - fixes some common problems in old (3.x and earlier) PRISM models
    - prism-statra - combines a PRISM .sta and .tra file into one file
    (these scripts run on Linux/Unix/OS X and require prism to be in your path)
+   - prism-filler.py - automates filling a text file (e.g. latex) with results
+     from log files. This script requires Python, run it without arguments for
+     more help.
+   - bash-prism-completion.sh - Bash programmable completion for PRISM,
+     automatically completes options for PRISM when run in Bash.
 
  * syntax-highlighters/ - Syntax highlighting configs for various tools
    - gnome/{Overrides.xml,prism.lang} - files for Gnome environment
diff --git a/prism/etc/scripts/prism-filler.py b/prism/etc/scripts/prism-filler.py
new file mode 100755
index 00000000..3b302de2
--- /dev/null
+++ b/prism/etc/scripts/prism-filler.py
@@ -0,0 +1,187 @@
+# coding=utf-8
+# run "python prism-filler.py" for more help
+
+import re
+import sys
+import os
+from optparse import OptionParser
+
+list_of_files = []
+directories = []
+
+def treat_file_names(file) : 
+	global list_of_files
+	#first get the actual string if a number or None is provided
+	if file is None : 
+		file = list_of_files[-1]
+	if type(file) is int :
+		file = list_of_files[file]
+	list_of_files.append(file)
+
+	#now get a path
+	#first try the directories
+	for d in directories :
+		joined = os.path.join(d, file) 
+		if (os.path.isfile(joined)) :
+			return joined
+	#otherwise just return the file itself (note: if it does not exist, an exception will be thrown anyway)
+	return file
+
+#extract total computation time
+def p_time(file = None) :
+	file = treat_file_names(file)
+	log_content = open(file).read()
+	matched = re.search("Time for model checking:[^0-9]*(?P<time>[0-9.]*)", log_content)
+	if matched is None :
+		return None
+	else :
+		return matched.group("time");
+
+#extract total number of states
+def p_states(file = None) :
+	file = treat_file_names(file)
+	log_content = open(file).read()
+	matched = re.search("States:[^0-9]*(?P<states>[0-9.]*)", log_content)
+	if matched is None :
+		return None
+	else :
+		return matched.group("states");
+
+#extract time for LP solving (TODO does not take setup into consideration)
+def p_lptime(file = None) :
+	file = treat_file_names(file)
+	log_content = open(file).read()
+	matched = re.search("LP problem solved in [^0-9]*(?P<time>[0-9.]*)", log_content)
+	if matched is None :
+		return None
+	else :
+		return matched.group("time");
+
+#extract number of variables in LP
+def p_lpsize(file = None) :
+	file = treat_file_names(file)
+	log_content = open(file).read()
+	matched = re.search("Number of LP variables[^0-9]*(?P<vars>[0-9.]*)", log_content)
+	if matched is None :
+		return None
+	else :
+		return matched.group("vars");
+
+#extract time for value iteration (in multi obj)
+def p_vitime(file = None) :
+	file = treat_file_names(file)
+	log_content = open(file).read()
+	matched = re.search("The value iteration.s. took [^0-9]*(?P<time>[0-9.]*)", log_content)
+	if matched is None :
+		return None
+	else :
+		return matched.group("time");
+
+#extract number of objectives (in multi obj)
+def p_numobj(file = None) :
+	file = treat_file_names(file)
+	log_content = open(file).read()
+	matched = re.search("Number of objectives: [^0-9]*(?P<num>[0-9.]*)", log_content)
+	if matched is None :
+		return None
+	else :
+		return matched.group("num");
+
+#extract number of weights used (in multi obj)
+def p_weights(file = None) :
+	file = treat_file_names(file)
+	log_content = open(file).read()
+	matched = re.search("Number of weight vectors used: [^0-9]*(?P<num>[0-9.]*)", log_content)
+	if matched is None :
+		return None
+	else :
+		return matched.group("num");
+
+#The program starts here
+descrip = u"A script that allows to fill in data from prism logs to textual files\
+ such as Latex or CSV.\
+ Usefull when writing papers with tables containing results of experiments.\
+ Basically, it searches the textual file for occurences of a pound sign (\u00A3), \
+ interprets the text between any two pound signs as a python program, executes\
+ it and replaces it with the output of the program (removing the pound signs). \
+ Some useful functions are provided to extract information from prism logs, see end of this help.\
+ Each of the functions takes one optional argument, which is a file name, \
+ or a number n which says that the n-th log file searched should be used \
+ (repetitions do count, and negative numbers are allowed). \
+ For example, 1 stands for the second file used, -2 stands for last but one file. \
+ If the argument is omitted, '-1' is used (i.e. last file)."
+ep = "Predefined functions: p_time (total computation time), p_states (number of states of model), \
+ p_lptime (time taken by LP solver),  p_lpsize (number of variables in linear program),\
+ p_vitime (total number for value iterations, in multi-objective queries), p_numobj (number of objectives, \
+ in multi-objective queries), p_weights (number of different weights used, in multi objective queries)."
+
+parser = OptionParser(usage="usage: %prog [options] input_file",version="alpha1 (2012-03-04)",description=descrip, epilog=ep)
+parser.add_option("-o", "--output", action="store", type="string", dest="outputfile", default=None, help="Output to file (stdout is used if not present")
+parser.add_option("-d", "--dir", action="store", type="string", dest="dir", default=None, help="(NOT SUPPORTED YET) Use directory to look for log files. Multiple directories are separated with colon ':', and the first file found is used.")
+parser.add_option("-i", "--ignore-errors", action="store_true", dest="ignoreerrors", default=False, help="The execution will not stop on the first error and the output file will be produced despite errors, with \"ERROR\" strings in it.")
+(options, args) = parser.parse_args()
+
+if len(args) != 1:
+	parser.print_help()
+	sys.exit(1)
+
+#open input file
+try :
+	file_content = open(args[0]).read()
+except IOError as e:
+	print "Cannot read the input file '" + e.filename + "': " + e.strerror >> sys.stderr
+	sys.exit(1)
+
+#open output file
+try :
+	if (options.outputfile is None) :
+		output_file = sys.stdout
+	else :
+		output_file = open(options.outputfile, "w")
+except IOError as e:
+	print >> sys.stderr, "Cannot access the output file '" + e.filename + "': " + e.strerror
+	sys.exit(1)
+
+#parse directories
+if not options.dir is None :
+	directories = options.dir.split(':')
+
+#do the work, first, get a string between two pound signs
+matched = re.search("£(?P<cmd>[^£]*)£", file_content)
+while not matched is None :
+	was_successful = True
+	#evaluate the user provided expression
+	replacement = None
+	try :
+		replacement =  eval(matched.group("cmd"))
+	except NameError as e :
+		print >> sys.stderr, "Cannot evaluate '" + matched.group("cmd") + "': " + e.args[0]
+		was_successful = False
+	except IOError as (noerror, strerror) :
+		print >> sys.stderr, "Cannot evaluate '" + matched.group("cmd") + "': " + strerror
+		was_successful = False
+	except IndexError as e :
+		print >> sys.stderr, "Cannot evaluate '" + matched.group("cmd") + "', you are probably trying to open n-th last file when less than n files were opened so far."
+		print >> sys.stderr, "The error message reported is: " + e.args[0]
+		was_successful = False
+	except SyntaxError as e :
+		print >> sys.stderr, "Cannot evaluate '" + matched.group("cmd") + "', invalid syntax. Have you enclosed the file name in qutation marks?"
+		print >> sys.stderr, "The error message reported is: " + e.args[0]
+		was_successful = False
+
+	if replacement is None and was_successful :
+		print  >> sys.stderr, "Evaluation of '" + matched.group("cmd") + "' did not return any text. Possible problems are using wrong log file or calling a wrong function."
+		was_successful = False;
+
+	#if there was some problem, either substitute with a dummy string or exit, depending on user preferences.
+	if (not was_successful) and options.ignoreerrors:
+		replacement = "ERROR"
+	if (not was_successful) and not options.ignoreerrors:
+		sys.exit(1)
+
+	#replace the text
+	file_content = file_content.replace(matched.group(),replacement, 1);
+	#get next match
+	matched = re.search("£(?P<cmd>[^£]*)£", file_content)
+#output the string with substitutions
+print >> output_file, file_content